It's hard to tell for sure because there are a few possible setups you can have with what you described. I would guess though that Server A is your cluster master server (i.e. the cluster is configured in the QEMC of Server A). Server B is most likely the cluster slave node. As a rule of thumb, I usually assume that whichever server has more tasks running is the master. But of course that is not always the case.
Either Server B or Server C handles Publisher tasks, because those are the only servers where Distribution Service is running. You might also have a either a dual or a clustered Publisher environment, in which case both Server B and C are Publishing.